SNP, GWAS, and post-GWAS

Identify disease genes

• Identifying genes that contribute to disease risk is one of the main objectives of molecular research.

• Such findings have contributed to improvements in diagnosis, prognosis, and therapy.

• With the successful identification of disease genes for many single-gene disorders, the focus has shifted to diseases with a complex, multifactorial etiology.

Genome wide association studies (GWAS)

• Identify associations between genetic variations (loci) and traits (including diseases).

  • Test for differences in the frequency of genetic variants between individuals who are ancestrally similar but differ phenotypically.

• Genetic variations:

  • Single nucleotide polymorphism (SNP) → most common.

  • Copy number variants.

  • Large sequence variations.

• Traits:

  • Disease (e.g., cancer, Alzheimer’s).

  • Phenotypes (e.g., height, eye color).

  • Gene expression (eQLT).

  • Etc.

Sinlge nucleotide polymorphism (SNP)

• Most common type of genetic variation:

  • Single base pair.

  • Responsible for 90% of all human genetic variations.

• SNPs are scattered all around the genome:

  • Within genes (coding SNPs): located within the coding region of a gene, may change the amino acid sequence of the gene’s protein product.

  • Outside of genes (non-coding, the majority): may change the timing, location, or gene expression level.

• What is a SNP:

  • A SNP occurs when another nucleotide replaces one nucleotide at a specific base pair position in the DNA sequence.

    • It can be a substitution, insertion, or deletion.

SNP database

• NCBI Short Genetic Variation database (dbSNP) catalogs short variations in nucleotide sequences for humans.

Major and minor allele

• Major allele is the most common variant found in a population, while the minor allele is the less common or rarer variant at that same position.

  • Minor allele frequency (MAF).

  • MAF > 1% → common SNP.

  • MAF < 1% → rare SNP.

• Major and minor alleles are population-specific.

• Focus often on the minor allele because minor alleles are crucial for identifying disease risks and studying genetic selection.

Synonymous vs. nonsynonymous SNPs

• Synonymous SNPs:

  • Do not change the amino acid sequence of a protein.

  • Silent change.

• Nonsynonymous SNPs:

  • Potentially alter protein structure and function.

• This distinction arises because the genetic code is redundant, with multiple codons sometimes coding for the same amino acid.

Missense vs. nonsense SNPs

• Missense and nonsense SNPs are both nonsynonymous mutations.

  • Change the amino acid sequence.

• Missense mutation:

  • Results in a different amino acid being incorporated into the protein.

  • It will alter the protein’s structure and function.

• Nonsense mutation:

  • This changes from a sense codon to a premature stop codon.

  • This leads to a truncated and often non-functional protein.

Linkage disequilibirum (LD) blocks

• Sets of nearby SNPs on the same chromosome are inherited together in blocks.

• Haplotype/LD block:

  • A group of alleles that are co-inherited as a single block.

  • LDTools: LDHap, LDMatrix, etc.

Tag SNPs

• A few SNPs are enough to identify the haplotypes in a block uniquely.

• Reduce the number of SNPs required to examine the entire genome for association with a phenotype.

Methods other than measure SNPs

• SNPs can be measured through genotyping and DNA sequencing.

• Genotyping examines a specific set of targeted sites within the genome using methods like SNP arrays.

  • A targeted approach, focusing only on pre-selected, known SNPs at specific positions in the genome.

• DNA sequencing reads the entire DNA sequence, allowing for the discovery of known and unknown SNPs.

  • SNPs captured through variant calling after sequence alignment → vcf files.

  • Allow for the discovery of novel or unexpected SNPs that are not known beforehand.

  • Read the entire DNA segment → more data and more expensive than genotyping.

SNP selection (genotyping)

• Identify DNA regions of interest.

• Identify patterns of SNPs that are inherited together on a chromosome.

  • HapMap project: tested the association of millions of SNPs across multiple populations → produced the SNP panels that are used today.

• Select tag SNPs that can represent a block of associated SNPs.

SNP genotype

• SNP genotypes can be coded differently based on genetic models.

  • Additive model (ADD) → commonly used.

  • Dominant model (DOM).

  • Recessive model (REC).

• SNP genotypes are commonly coded as 0, 1, or 2 to represent the number of copies of the minor allele.

  • 0 → homozygous dominant (e.g., AA).

  • 1 → heterozygous (e.g., Aa).

  • 2 → homozygous recessive (e.g., aa).

Genotype/haplotype phasing

• Current technology gives genotypes but not haplotypes.

• The process of determining which alleles are located on the same chromosome, i.e., the haplotype.

• Phasing tools: SHAPEIT5, BEAGLE5.5.

GWAS association testing

• A simple statistical test, such as chi-square.

  • Case control analysis.

  • Without confounding factors.

• Linear regression models.

  • If the phenotype is continuous (such as height, blood pressure, or body mass index).

• Logistic regression models.

  • If the phenotype is binary (such as the presence or absence of disease).

• Linear mixed models.

  • Can account for genetic relatedness among individuals.

GWAS example

Odds ratio

• Measure of effect size.

• OR = 1, no disease association.

• OR > 1, allele C increases risk of disease.

• OR < 1, allele C decreases the risk of disease.

PLINK

• Mostly shared.

GWAS results visualized

• Manhattan plot:

  • Bonferroni testing threshold of p < 5 × 10^-8.

    • Multiple test correction.

  • Genome-wide significance threshold = red line on the image.

    • The peaks are SNPs that are close to each other = co-inherited.

  • Significance threshold varies depending on:

    • Number of SNPs examined.

    • Number of subjects included.

    • Minor allele frequencies.

    • Etc.

• LocusZoom: a combination of a Manhattan plot and a genome browser.

Genotype datasets for large scale GWAS

• Biobanks and large population-based studies with genetic and phenotype data available for research.

  • For USA → ‘All of Us’ initiative or 23andMe.

• Genotype data are typically restricted due to re-identification risk.

  • Application needed.

Databases for GWAS summary statistics

• GWAS Catalog.

• GWAS Atlas.

• Allow easy access to summary statistics for thousands of traits.

• Many downstream analyses are built on the summary statistics, rather than the raw genotype itself.

Post-GWAS analysis

• Functional mapping:

  • Where are they located in the DNA?

  • Pathways enriched by GWAS findings.

  • How may they influence molecular functions leading to disease?

  • Identify the tissues or cell types where these variants are likely to act.

• Causal variants identification:

  • GWAS findings come up as clusters → highly correlated.

  • Which one is likely causal?